The safety of a surgical patient is closely related to the quality of the surgeon's training. This is particularly true of patients undergoing cardiac surgery, where stakes are high and technical skills are demanding. In most surgical training, technical skills are taught by apprenticeship: residents learn surgery in the operating room, doing parts or all of real operations on real patients. Unfortunately, today's operating room provides insufficient time in which to teach surgery, has no tolerance for the inefficiency inherent in education, offers no chance for deliberate practice of skills, and cannot provide for orchestrated training in how to deal with adverse events. But all of these are essential to the training of a safe surgeon. This study intends to show that training in cardiac surgical techniques can be improved by using modern cardiac surgery simulation technology combined with a rigorous, simulation-based curriculum. This should produce surgeons with better skills and thus ensure safety for patients. This three-year project will determine the effectiveness of using simulator-based training in component tasks and overall procedures based on six modules: three types of cardiac surgical operations and three significant adverse events that can occur during cardiac surgery. The procedures will be taught using a computer-controlled, tissue-based cardiac surgery simulator which has been shown to realistically duplicate the actual patient undergoing cardiac surgery. Eight institutions (University of North Carolina at Chapel Hill, Massachusetts General Hospital, Johns Hopkins University, Vanderbilt University, University of Rochester, Mayo Clinic, Stanford University, and University of Washington) will participate in the study, which will include 16 first-year cardiothoracic residents in each of two consecutive academic years. The first phase of the project (six months) will include development of the simulators and a comprehensive curriculum, and faculty education. The second phase (24 months) will involve simulator-based training and performance assessment of cardiac surgery residents. Each part of the simulation-based training will be evaluated by the instructor, using approved assessment tools. At the end of the training, two independent cardiothoracic surgeons will grade videotaped surgery performed by each simulator-trained resident plus surgery by 16 reference residents. Assessment data from each site will be entered into a study database which will be developed and managed at the University of North Carolina. In the last phase (six months), data will be analyzed by standard methods of analysis. While this proposal will test the hypothesis that cardiac surgery residents can be trained to be safer surgeons by using appropriate simulator-based training, the results should apply equally well across a broad spectrum of surgical practice.